Years ago, I was staying with a few flatmates during an exchange program. Conversations in this type of setting tend to have a good dose of randomness, and here is a funny one that I remembered while writing this post.

One day, one of my flatmates asked: “Have you seen a red and purple toothbrush?” As it turns out, I have seen it, especially with that eye-catching color scheme, but couldn’t fully remember where. As any person would, I just guessed and replied, “Yeah, have you checked in the bathroom?” to which they replied, “Seems it’s not there.” “Then I don’t know, sorry”, I replied.

Later during the day, I noticed that said toothbrush was near the kitchen sink, and it hit me: that’s where I had originally seen it, and in fact, I remembered seeing this mundane object because I had seen it in this odd location.

To be honest, I think I wouldn’t have ever recalled this anecdote if not because the article we are covering actually used an experimental design consisting of letting people study items presented in expected or unexpected places, which happened to be kitchen or bathroom items presented in either a kitchen or bathroom context!

In Manipulating belief partially remedies the metamemory expectancy illusion in schema-based source monitoring, published in Memory & Cognition, Marie Luisa Schaper and Ute J. Bayen (pictured below) present a sophisticated study where they tested whether a procedure to manipulate a commonly held belief that often leads to metamemory illusions can prevent the illusion.

If this sounds complex, it is because it is, so don’t let yourself be misled by the simplicity of the opening anecdote. And yet, the methodology created by the authors allowed the use of a relatively simple experimental approach to disentangle a variety of theoretical and empirical contributions to metamemory illusions. So, let’s break it down together, shall we?

Metamemory illusions and ways to remedy them

Before we recall something, such as a piece of information or the context in which we learned about it, we may make predictions about our future memory. This is an example of metamemory, and there are many variables that can influence it.

For example, when we study a certain material, our fluency in processing it and our beliefs about how memory works can affect our metamemory judgements. For example, if we feel that a certain material is easy to study (i.e. high processing fluency), or if we believe that we will be more likely to remember the source of a study material if it is presented in an expected, vs an unexpected context (a common belief about our own memory), we are more likely to predict that we will recall the material better (a metamemory judgement).

However, if we base our metamemory judgements on invalid predictors of our actual memory performance, a metamemory illusion arises. In other words, we may over- or under-estimate the contributions of experiential and belief factors in our actual memory performance.

Accordingly, if one wants to remedy those illusions, an approach could be to manipulate experiential or belief aspects associated with those illusions. The authors of the study focused on testing whether a manipulation designed to correct an often-observed incorrect belief (i.e., that we will better remember the context in which we studied items when the context is expected vs. unexpected) affects individuals’ metamemory judgements.

Toothbrushes in the kitchen and frying pans in the bathroom

The authors conducted two experiments in which they let participants study items found typically in either a bathroom (e.g., a toothbrush) or a kitchen (e.g., a frying pan) presented in an expected (toothbrush in the bathroom) or unexpected (toothbrush in the kitchen) context, and then tested how well they recalled items and their sources.

Participants in the experimental group of each experiment started it with a manipulation designed to correct often-held incorrect beliefs: that a high subjective feeling of processing fluency, and that studying items in an expected (vs unexpected) context, would make it more likely for them to remember the items and the contexts where they were presented. In contrast, participants in the control groups were not exposed to this belief-correcting manipulation and were therefore likely to hold the incorrect beliefs.

The experimenters also measured participants’ judgements of how likely they were to remember the context (bathroom or kitchen) in which they had studied each item, using a scale from 0% to 100% (a measure called Judgement of Source – JOS).

Lastly, the experimenters also measured participants’ beliefs about the percentage of contexts they recalled over the whole experiment. This was done at the end of both experiments, and for experiment 2 only, it was also included as a prediction measure before the start of the study phase.

Correcting and utilizing beliefs to render metamemory judgements

In both experiments, the groups of participants exposed to the belief-correcting manipulation decreased (but did not fully eliminate) the degree to which they showed the metamemory illusion of expecting to recall expected sources for studied items better than for unexpected sources. In other words, while they actually showed better memory for unexpected than expected sources, they believed that they had a better memory for expected than unexpected sources. Figure 2 illustrates participants’ beliefs before (for Experiment 2) and after the test phase (for Experiments 1 and 2).

A closer look at the data (and the additional measures included in Experiment 2) allowed to determine that the partial persistence of the metamemory illusion was based on several factors: some participants may have failed to correct their beliefs (i.e. come to believe that memory for unexpected sources is better than for expected sources after the manipulation); and/or to utilize their corrected beliefs (i.e. believing that memory for unexpected sources is better than for expected sources but not being able to apply this belief to the judgements about their own memory). In addition, evidence of individual differences in the extent to which participants showed those deficits in correcting or using their beliefs was found.

According to the authors,

“Our study shows that correcting false beliefs about memory can reduce—but not fully eliminate—a common metacognitive illusion: the mistaken belief that we remember expected information sources better. This suggests that both belief and experience shape our predictions about memory, with implications for learning, legal decisions, and misinformation detection.”

As for the opening anecdote, why was my flatmate’s toothbrush in the kitchen? I mean, in experimental contexts, like in movies, we often get exposed to unexpected situations, but here? It turns out that someone was taking a shower, and so my flatmate thought it was easier to brush their teeth in the kitchen sink rather than wait, as they were already in a hurry. It was a practical solution, I would concede, even if not necessarily something I would do. In any case, as unexpected as this was, in real life I would rather have seen the toothbrush in the kitchen than a frying pan in the toilet, for which I cannot fathom a plausible reason to be in that unexpected context.

Psychonomic Society’s article featured in this post:

Schaper, M.L. & Bayen, U.J. (2025). Manipulating belief partially remedies the metamemory expectancy illusion in schema-based source monitoring. Memory & Cognition https://doi.org/10.3758/s13421-025-01757-2